AIM: To evaluate the usefulness of stool-PCR test for diagnosis of Helicobacter pylori (H pylori) infection in pediatric populations.

METHODS: Based on endoscopic features (including nodular gastritis, erosive duodenitis and ulcer) and/or a positive rapid urease test (RUT) obtained during endoscopy, 28 children from a group of children admitted to the Children's Medical Center of Tehran for persistent upper gastrointestinal problems were selected to compare biopsy-based tests with stool-PCR. Their gastric activity and bacterial density were graded by the updated Sydney system, and their first stool after endoscopy was stored at -70°C. Biopsies were cultured on modified campy-blood agar plates and identified by gram-staining, biochemical tests, and PCR. Two methods of phenol-chloroform and boiling were used for DNA extraction from H pylori isolates. Isolation of DNA from stool was performed using a stool DNA extraction kit (Bioneer Inc, Korea). PCR was performed using primers for detection of vacA, cagA, and 16srRNA genes in both isolates and stool.

RESULTS: Sixteen out of 28 child patients (57%) were classified as H pylori positive by biopsy-based tests, of which 11 (39%) were also positive by stool-PCR. Sensitivity and specificity of stool-PCR was 62.5% and 92.3% respectively. H pylori was observed in histological sections for 10 out of 11 stool-positive patients. Association was observed between higher score of H pylori in histology and positivity of stool-PCR. Also association was observed between the more severe form of gastritis and a positive stool-PCR.

CONCLUSION: Association between higher score of H pylori in histology and a positive stool-PCR make it a very useful test for detection of H pylori active infection in children. We also suggest that a simple stool-PCR method can be a useful test for detection of H pylori virulence genes in stool.

Helicobacter pylori (H pylori) infection in humans is associated with gastritis, gastric ulcer, and gastric cancers[12]. Infection occurs mainly in childhood and infected individuals usually carry it for life unless treated[34]. Epidemiology of infection by H pylori has been characterized by a linear increase with age in western industrial countries and by a large number of children and juveniles being infected in developing countries[5]. Currently used methods for diagnosis of H pylori infection, such as culture, histology, and rapid urease test (RUT) are very sensitive and highly specific tests, but require invasive sampling. The non-invasive methods, such as serology and urea breath test (UBT), are also sensitive and specific; however, positive results obtained by serology do not necessarily indicate current infection by H pylori[67]. UBT requires an expensive instrument, which is not always available in routine clinical laboratories, especially in developing countries. In addition the performance of the test has been associated with some disadvantages for infants and very young children, as well as patients with certain neurological disorders[67]. H pylori is not an intestinal pathogen, and therefore is expected to be present in low concentrations in stool; however, it can be detected in stool specimens by H pylori stool-antigen (HpSA) test, PCR, or even culture[89101112]. The HpSA test has been shown to be very useful, especially in children; however, various commercial tests have shown some discrepancies in different geographical areas[131415]. Stool-culture is a very specific method; however, the massive numbers of diverse micro-organisms in stool makes it very difficult in routine practice[812]. Stool-PCR may also be a very useful method in detection of H pylori infection, but reported success rates for the detection of H pylori DNA in feces vary from 25% to 100%[68]. This variability is probably due to H pylori degradation in the gastrointestinal tract and/or the presence of inhibitors such as complex polysaccharides[1617]. The purpose of this study was to evaluate the usefulness of the stool-PCR test for diagnosis of H pylori infection in pediatric populations.

In our previous study[12], we successfully cultured H pylori from stool; however, the sensitivity of stool-culture was low. Using PCR, we detected H pylori specific genes in isolates and stool in sick and healthy children. However, when fecal extracts were not subjected to column chromatography, there were no results even for the positive controls. This suggests that the method of DNA extraction used in this work efficiently removed the PCR inhibitors. Various methods has been used for the removing of inhibitors or for the purification of DNA such as the removal of PCR inhibitors by a polypropylene filter, dilution of fecal suspension, and DNA purification by various biochemical techniques; in many studies with filtration of stool and column chromatography, high sensitivity was observed[8101114222324].

In this work, by detection of various H pylori specific genes in stools, 62.5% sensitivity and 92.3% specificity was observed for stool-PCR (Table 2). Nevertheless, by PCR only one or two out of three H pylori specific genes were detectable (Table 3). While this permits us to think that the absence of amplification is related to the absence of the detecting gene from the genome or the absence of intact template DNA (in stool), it would be a premature conclusion, since PCR-based absence of an ORF does not necessarily mean its absence from the genome. Also, in a highly recombining genome like

H pylori, PCR primer annealing sites can pose problems and amplifications may not be generated[2526]. Thus, we think that for genotyping of H pylori from stool, using more than one primer for each gene may enhance detection rate. Many investigators have proposed semi-nested or nested PCR as more sensitive methods for stool-PCR[810]. Although these methods reduce background, their disadvantages would be presence of false positive results due to detection of dead bacteria in stool even in low amounts. Sensitivity and specificity of stool-PCR method in this work were acceptable, suggesting that PCR method used in this work was quite adequate for this evaluation.

H pylori is not an intestinal pathogen, and therefore is expected to be present in low concentrations in stool; however, the status of the infection of H pylori may influence its density in stool. Thus, we compared histological scoring of H pylori with pathological grading and also with the results of stool-PCR. Concordance was observed between higher score of H pylori in histological sections and a positive stool-PCR (Table 4). Also, association was observed between the more severe form of gastritis and a positive stool-PCR. Therefore, the degree of stomach colonization by H pylori may be important for successful detection of DNA in stool samples. Otherwise, the amount of bacteria excreted in stool may reveal information on the status of H pylori infection. Consequently, the association between a higher score of H pylori in histology and a positive stool-PCR make it a very useful test for detection of pediatric H pylori infection.

We thank MRS Shahla Tafreshi and MRS Masoumeh Madadi for technical assistance. Mavenat Pajouheshy of Alzahra University, Vanak, Tehran has supported this work.